On Fri, Jul 31, 2015 at 06:10:37PM +0200, 'Klaus Aehlig' via ganeti-devel wrote:
The 7th shell test for hspace relied on unspecified
behavior: tiered allocatio requires the resource causing
the largest number of allocations to be immpossible to
be shrunk first, if that leads to a valid allocation.
However, in this example, the actual failure statistics
is
[(FailMem,2),(FailDisk,0),(FailCPU,0),(FailN1,0),(FailTags,0)
,(FailMig,0),(FailDiskCount,0),(FailSpindles,2),(FailInternal,0)]
So it is up to the implementation to choose on whether to shrink
spindles or memory first. Therefore, accept the results of both
shrinking orders.
Signed-off-by: Klaus Aehlig <[email protected]>
---
test/hs/shelltests/htools-hspace.test | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/test/hs/shelltests/htools-hspace.test
b/test/hs/shelltests/htools-hspace.test
index 1ff0680..9e3da7b 100644
--- a/test/hs/shelltests/htools-hspace.test
+++ b/test/hs/shelltests/htools-hspace.test
@@ -26,7 +26,7 @@
>>>=0
# Mixed cluster, half with exclusive storage
-./test/hs/hspace --machine-readable -t $TESTDATA_DIR/hspace-tiered-mixed.data --no-capacity-checks > $T/capacity && sh -c ". $T/capacity && test
\"\${HTS_TSPEC}\" = '131072,1048576,4,12=2 131072,1048576,4,10=2 129984,1048320,4,10=2' && test \"\${HTS_ALLOC_INSTANCES}\" = 6 && test
\"\${HTS_TRL_SPN_FREE}\" = 0 && test \"\${HTS_FIN_SPN_FREE}\" = 18"
+./test/hs/hspace --machine-readable -t $TESTDATA_DIR/hspace-tiered-mixed.data --no-capacity-checks > $T/capacity && sh -c ". $T/capacity && echo
\"\${HTS_TSPEC}\" | grep -q '131072,1048576,4,12=2 .*129984,1048320,4,10=2' && test \"\${HTS_ALLOC_INSTANCES}\" = 6 && test
\"\${HTS_TRL_SPN_FREE}\" = 0 && test \"\${HTS_FIN_SPN_FREE}\" = 18"
>>>=0
# Verify that instance policy for disks is adhered to
--
2.5.0.rc2.392.g76e840b
LGTM, thanks
